Journal of Research in Music Education-2011-Norgaard-109-27

8/16/2019 Journal of Research in Music Education-2011-Norgaard-109-27

http://slidepdf.com/reader/full/journal-of-research-in-music-education-2011-norgaard-109-27 1/20

http://jrm.sagepub.com/ Music Education

Journal of Research in

http://jrm.sagepub.com/content/59/2/109The online version of this article can be found at:

DOI: 10.1177/0022429411405669

2011 59: 109Journal of Research in Music Education

Martin NorgaardDescriptions of Improvisational Thinking by Artist-Level Jazz Musicians

Published by:

http://www.sagepublications.com

On behalf of:

MENC: The National Association for Music Education

can be found at:Journal of Research in Music Education Additional services and information for

http://jrm.sagepub.com/cgi/alertsEmail Alerts:

http://jrm.sagepub.com/subscriptionsSubscriptions:

http://www.sagepub.com/journalsReprints.navReprints:

http://www.sagepub.com/journalsPermissions.navPermissions:

http://jrm.sagepub.com/content/59/2/109.refs.htmlCitations:

by Bryan Chesi on August 24, 2011 jrm.sagepub.comDownloaded from

http://jrm.sagepub.com/


http://jrm.sagepub.com/content/59/2/109



http://www.sagepublications.com/

http://www.menc.org/

http://jrm.sagepub.com/cgi/alerts


http://jrm.sagepub.com/subscriptions

http://www.sagepub.com/journalsReprints.nav


http://www.sagepub.com/journalsPermissions.nav

http://jrm.sagepub.com/content/59/2/109.refs.html






http://www.sagepub.com/journalsPermissions.nav


http://jrm.sagepub.com/subscriptions


http://www.menc.org/

http://www.sagepublications.com/





Journal of Research in Music Education

59(2) 109 –127

© 2011 MENC: The NationalAssociation for Music Education

Reprints and permission:

sagepub.com/journalsPermissions.navDOI: 10.1177/0022429411405669

http://jrme.sagepub.com

JRM405669 JRMe59210.1177/0022429411405669NorgaardJournal of Research in MusicEducation

1Georgia State University, Atlanta, GA, USA

Corresponding Author:

Martin Norgaard, School of Music, Georgia State University, P.O. Box 4097, Atlanta, GA 30302-4097

Email: [email protected]

Descriptions of

Improvisational Thinking by

Artist-Level Jazz Musicians

Martin Norgaard1

Abstract

Thought processes of seven artist-level jazz musicians, each of whom recorded animprovised solo, were investigated. Immediately after completing their improvisations,participants listened to recordings of their playing and looked at the notation of theirsolos as they described in a directed interview the thinking processes that led to therealization of their improvisations. In all of the interviews, artists described making sketchplans, which outlined one or more musical features of upcoming passages. The artistsalso described monitoring and evaluating their own output as they performed, making judgments that often were incorporated into future planning. Four strategies used bythe artists for generating the note content of the improvisations emerged from the

analysis: recalling well-learned ideas from memory and inserting them into the ongoingimprovisation, choosing notes based on a harmonic priority, choosing notes based on amelodic priority, and repeating material played in earlier sections of the improvisation.

Keywords

improvisation, jazz performance, cognition, jazz education

The act of musical improvisation is central to the jazz idiom (Schuller, 1968). Although jazz music comprises a number of stylistic conventions that distinguish it from other

musics, it is the improvisatory element in jazz—the development and expression of musi-

cal ideas in the moment—that has defined jazz throughout its history (Gridley, 1987).

In the earliest incarnations of instrumental jazz, musicians learned to improvise

through listening, imitation, and working alongside masters of the art form (Berliner,

1994). More recent initiates to jazz in America have learned to improvise through more







110 Journal of Research in Music Education 59(2)

formalized programs of jazz instruction. In addition to college jazz programs, many

musicians are introduced to jazz improvisation through high school jazz bands and

introductory college courses for nonjazz majors (Azzara, 2002). Consequently, the

responsibility of continuing the tradition of jazz improvisation has shifted to the educa-tor, which raises the question of how best to design authentic learning experiences.

Designing meaningful instruction for novices in any discipline is informed by under-

standing the thought processes of advanced practitioners (Bruner, 1977; Duke, 2005).

But what is involved in the thinking of artist-level improvisers? If the study of tech-

nique and music theory fails to develop some components of this thinking, as some

have suggested (Berliner, 1994; Kenny & Gellrich, 2002; Pressing, 1988; Sarath, 2002),

is it possible to identify these components by studying artist-level improvisers?

The innumerable momentary decisions made by jazz improvisers are influenced by

multiple factors. Note choices are guided by the accompanying chord structure andrhythmic feel and are influenced by preceding events and intended goals (Pressing,

1984). Rhythmic feel affects note placement, duration, melodic shape, and inflection.

Interactions with other ensemble members add further complexity to the decision-

making process.

The thought process guiding tonal jazz improvisation has been compared to the

thinking that supports spoken language, because both are created in real time (Berkowitz,

2010; Berliner, 1994; Pressing, 1998). Word choice is guided by syntactic rules just as

tonal improvisation is guided by melodic, rhythmic, and harmonic rules (Patel, 2003).

Language production may pass through several stages in which an idea behind a phrase first is planned, then translated into linguistic structure, executed to create overt

speech, and finally evaluated through a monitoring process (Levelt, Roelofs, & Meyer,

1999). Similarly, an improviser may conceptualize a musical goal initially, then formu-

late the goal according to the key and type of tune, decide upon and execute a motor

plan, and finally evaluate the result (Berkowitz, 2010). Language also contains words

and combinations of words that are part of a speaker’s vocabulary (Harley, 2008;

Levelt et al., 1999). Similarly, improvisers accrue a collection of melodic figures that they

incorporate into their improvisations (Berkowitz, 2010; Berliner, 1994; Finkelman,

1997; Owens, 1995; Pressing, 1998). At times, these figures are inserted into theimprovisation note-for-note (Gushee, 1991). At other times, the improviser may use

various procedures to create musical figures similar to those stored in long-term mem-

ory yet presented with alterations (Finkelman, 1997).

Speculative models have been proposed to explain how rules and vocabulary affect

improvisational decision making (Clarke, 1988; Johnson-Laird, 2002; Kenny &

Gellrich, 2002; Pressing, 1988). In one model, Pressing (1988) divided improvisations

into collections of note groupings that he labeled “events.” Each event is triggered by

a creative intention in the form of a mental schema that contains a cognitive image of

sound and corresponding motor realization. As the mental schema is executed, theimproviser compares the intention with the actual performed output through various

feedback loops. New events often share features with preceding events, resulting in a







Norgaard 111

related set of note groupings. Yet, at times, the improviser may choose to interrupt the

flow by initiating an event that is completely unrelated to the preceding event.

Johnson-Laird (2002) argued that note groupings are less important and instead

focused on learned rules that dictate individual note choices as the main guiding prin-ciple of improvisation. He proposed that advanced improvisers have internalized

so completely these rules that choices are made implicitly and with minimal use of

working memory. This would leave the conscious mind free to focus on architectural

aspects of the solo and interaction with other ensemble members. This model predicts

that practitioners would describe focusing on structural considerations while individ-

ual note choices are made outside of consciousness.

One recent study using functional magnetic resonance imaging (fMRI) appears to

support the notion that the improviser is not controlling all improvisational choices

consciously (Limb & Braun, 2008). Limb and Braun compared the neural activationsof expert jazz pianists while they improvised and while they performed memorized

music. They found that the brain regions related to conscious decision making and

self-monitoring (lateral orbitofrontal and dorsolateral prefrontal cortices) were less

active during improvisation than during performance of memorized passages. In con-

trast, increased activation was seen in an area known to control behaviors that conform

to internalized rules outside of conscious awareness (frontal polar portion of the medial

prefrontal cortex). This area is associated with “organizing internally motivated, self-

generated, and stimulus-independent behaviors” (p. 4). Limb and Braun’s results

appear to support Johnson-Laird’s (2002) notion that choices during improvisation aremade implicitly according to learned rules.

Artist-level thinking in jazz has been analyzed using qualitative methodologies

(Berliner, 1994; Hargreaves, Cork, & Setton, 1991; Mendonça & Wallace, 2004;

Monson, 1996). Berliner (1994) conducted a landmark study, based on interviews with

over 50 renowned jazz artists, exploring learning environments and various aspects of

improvisational thinking. The interviewees were asked questions about improvisation

most often without reference to specific music examples, however, and their answers

were seldom specific enough to illuminate their actual thought processes. Berliner

provided specific music examples in solo transcriptions that illustrate some of thegeneral findings of his study, locating the examples post hoc according to his interpre-

tations of his informants’ comments.

Another qualitative study by Monson (1996) focused primarily on the interactive

element of improvisation. In 14 interviews, Monson asked specific questions about

actual music examples from her informants’ extant recordings, but the material had

been recorded long before the interviews were conducted. Given the time between the

recorded performances and the interviews about them, it seems unlikely that the musi-

cians’ responses were accurate recollections of their thought processes at the moment

of creation. Only one recent study employed a protocol in which the participantsdescribed their improvisational thinking immediately following their performances

(Mendonça & Wallace, 2004). The focus of the study was to investigate the thinking








processes of musicians while improvising in a jazz duo in both tonal and atonal con-

texts. Analysis was done by assigning verbalizations to preset categories and compar-

ing the number of verbalizations in each category across the tonal and free conditions.

The participants spoke about their performances while listening and looking at videoand did not have access to notation of their improvisations. The data reported deal

primarily with the interactions between the two players in the duos.

A recently published study of cognition in classical improvisation combined quali-

tative research with historical analysis and neuroscientific data (Berkowitz, 2010).

Berkowitz analyzed several historical treatises on improvisation and conducted inter-

views with two prominent improvisers in the classical idiom. He considered these

findings in reference to brain imaging data that he collected as other classical pianists

improvised simple melodies. Berkowitz’s conclusions are in line with much of the infor-

mation reviewed above, including the performer’s use of prelearned fragments in animplicit process that is evaluated after execution.

The purpose of the present study was to describe the thinking processes underlying

expert jazz improvisation. The rationale for the investigation was premised on the

notion that descriptions of artist-level improvisers’ thinking may guide the develop-

ment of improvisational activities in the classroom that would be beneficial and grati-

fying to all learners.

Artist-level jazz improvisers were asked to improvise on a known chord progression

and were interviewed immediately afterward. I used as the basis for the interviews audio

recordings and approximate transcriptions of their improvisations, which I producedimmediately after their creation. I asked my informants to narrate their improvisational

thinking processes as they listened to and watched the notation of their just-completed

performances, linking comments to specific musical material. The limitations of previous

qualitative research concerning specificity and time between performance and interview

thereby were addressed by the current research design.

Method

Seven artist-level jazz improvisers participated in the study. Each artist had extensiveexperience in improvisation both in live settings and in the recording studio. All par-

ticipants had produced professional audio recordings; all were active performers with

busy performance careers in their local communities; and all had national and inter-

national touring experience. The participants performed on the following instruments:

acoustic bass, violin, trumpet, piano, guitar, alto saxophone, and trombone. At the

time of the study, the bassist, age 63, had been a mainstay of the East Coast jazz scene

since the 1970s and had served as the director of a major university jazz program for

20 years. The Grammy-nominated violinist, age 54, was part of the West Coast jazz

and bluegrass scene. The trumpeter, age 55, taught at a major Midwestern universityand was active on the jazz scene of Kansas City. The trombonist, 36, was part of a

nationally recognized big band and served on the faculty of a major university in

Canada. The pianist, age 46; guitarist, age 55; and saxophonist, age 53, were on the







Norgaard 113

faculty of a major Southwestern university. All the participants were male. Selection

of participants was based on availability and willingness to participate. The subject

matter of the interviews was personal in nature, requiring a level of openness from the

interviewees, who were not granted anonymity. The study protocol was approved bya university’s Institutional Review Board.

Each interview session had two major phases: the recording of the improvised solo

and the interview concerning the solo just performed. I asked participants to perform

a blues in F major using a melody of their choice, playing the melody first and then

improvising a solo. The improvisations were accompanied by only a drum track. No

particular version of the blues chord progression was specified, and no sheet music

was shown to the participants. I asked the participants to play as many choruses as they

wished before concluding their performances. The verbal directions read aloud to the

participants were as follows:

Play a blues in F major in a medium-up tempo. Play a melody you are very famil-

iar with followed by an improvisation on the blues form. Play long enough that

you feel the performance has a formal sense to it. In other words, go until you feel

like you can finish. You don’t have to decide the exact length beforehand.

All the performers in the study, except for the pianist, played along with a drum

accompaniment that was played through speakers during the performance. I created

the drum track accompaniment prior to the interviews by digitally looping two mea-sures of a swing drum track taken from an existing recording (Carman, 1999). The

performance tempo was 212 beats per minute.

Participants were recorded using a combination of software running on a laptop

computer. A Shure SM-81 microphone was connected to a USB audio interface, which

was connected to a laptop through the USB port. The audio recording software,

Samplitude 9 Professional (Tost, 2007), rendered the audio in a wave file format. In

the minutes following the recording phase of the session, I converted the audio file to

a MIDI file using TS-Audio To MIDI Realtime Converter (Egorov, 2004). I then

imported the MIDI file into the audio software, which allowed me to display the nota-tion in synchrony with the original audio recording.

Each session was captured on video. The camera was positioned with the partici-

pant in view during the performance and with the computer screen in view during the

interview.

During the interviews, I asked each participant to comment on the content of the

recorded solo as he listened to the audio and looked at the notation, and I directed him

to focus his verbal comments on aspects of decision making and structure. I read the

following at the beginning of the interview:

As you are watching and listening to your performance, try to narrate your con-

scious thinking, considering questions like, “Where did that come from?” We

are looking for a narration similar to a director’s commentary on a DVD. We








are particularly interested in how much of what you played comes from a reper-

toire or bank, how much is from a repertoire but modified in some way, and how

much is material you never played before that is generated in the moment. We

are also interested in the timing of decisions. Where are the decision points andwhat material does each decision affect?

I then played in sequence short sections of the improvised solo that corresponded

to musical phrases, pausing after each phrase. Participants often described their think-

ing without prompting. When necessary, I asked questions to clarify the participants’

comments using the guidelines of responsive interviewing (Rubin & Rubin, 2005),

which allow for follow-up questions to themes introduced by the interviewee and for

probes to keep the conversation on topic. Other than the instructions listed above, no

specific questions were prepared in advance.A preliminary session with the pianist was conducted prior to the other interviews

and was attended by the author and two additional researchers, all of whom asked

questions. The additional researchers both had considerable research expertise and

knowledge of jazz improvisation and served on the faculty of a major Southwestern

university. This preliminary session was used as a model for later interviews. The

musical task was selected and the verbal script used in later interviews was developed

from questions and comments that arose in this interview. As indicated previously, the

pianist did not play with a drum track. The tempo he chose was analyzed later and was

used to develop the accompaniment track for subsequent sessions. The data from the pianist’s interview were incorporated into the analysis.

I began the analysis of the interviews by transcribing all verbal information,

although I did not transcribe all utterances in detail. Concerning the level of detail

necessary in transcriptions of verbal interviews, Rubin and Rubin (2005) recommend

including only the details that are likely to be analyzed. I did not transcribe pauses,

repeated individual words, and utterances such as “hmm” and “ahh” if they did not

affect meaning.

The analysis of interview content was performed with the aid of the qualitative

analysis software ATLAS.ti 5.2 (Muhr, 2006), which has several advantages related toorganization and navigation as compared to manual indexing (Campbell, Connell, &

Beegle, 2007). I assigned a code to each phrase, sentence, or paragraph in my tran-

scriptions, according to the ideas expressed in each, using codes that emerged during

the analysis (Strauss & Corbin, 1998). I assigned a total of 121 codes to 563 quotations

within the seven interviews.

The process of identifying and coding conceptual categories has been described as

a creative, interpretive process (Charmaz, 2006; Glaser, 1998) in which no standard

form of analysis exists (Henwood & Pidgeon, 2003). To identify conceptual catego-

ries, the investigator compares “codes with codes (to generate a set of main themes)and then codes with data and data with data” in an interactive process (Henwood &

Pidgeon, 2003, p. 148). As I analyzed these categories, I realized that participants

often used different labels to describe the same phenomena. For example, the labels







Norgaard 115

initially attached to comments referring to the use of material from a bank of ideas

were “nuggets of stuff,” “units,” and “vocabulary dictated by history.” This process

included frequent meetings with the two senior researchers in which all of us dis-

cussed and finally agreed on the emerging conceptual categories.I also created exact transcriptions of all of the artists’ solos. The transcription soft-

ware used during the interview sessions yielded inexact transcriptions of the recorded

solos. In the final part of my analysis, I transcribed the improvised solos into fully

accurate music notation and linked participants’ most salient verbalizations to the

musical material with which they were associated. These transcriptions were checked

for accuracy by a graduate student majoring in jazz performance.

The comparison of two main types of data, verbalizations and musical output, and

the interpretations of three investigators constituted data triangulation in the current

study. Triangulation “is often used as a qualitative equivalent of validity and reliability”(Willis, 2007, p. 218). The six main themes discussed in the Results section were iden-

tified both through the qualitative analysis of verbalizations and in the improvised

musical output of the participants. Furthermore, the link between the verbal and musi-

cal output was supported by video data showing specifically where participants were

pointing in the transcription of the improvisation as they described their thinking.

Although the analysis model known as grounded theory specifies that the researcher

should not consider previous literature during the analysis to guide against bias (Glaser

& Strauss, 1967), in reality, “theory cannot simply emerge from or reflect data, because

interpretation and analysis is always conducted within some preexisting conceptualframework brought to the task by the analyst” (Henwood & Pidgeon, 2003, p. 134). It

is possible that extensive knowledge of the literature concerning improvisation and the

main researcher’s background in jazz performance and education influenced the anal-

ysis. In particular, the naming of the conceptual categories that serve as the main

themes in the results may have been influenced by prior knowledge.

Results

Six main themes emerged from the analysis of the qualitative information. Thesethemes reflect conceptual categories that emerged during my analysis. I determined

the main themes based on the frequency with which they appeared and their broad

applicability in describing the salient features of the artists’ performing and thinking.

The results include a description of two ongoing processes: a sketch planning process

and an evaluative monitoring process. These pertain to the forward-looking act of

planning and the backward-looking act of monitoring what has just been played. In

addition, I identified four strategies that served as the basis for the generation of musi-

cal material: the use of memorized music, which I refer to as an idea bank; the selec-

tion of notes with particular attention to the harmonic structure of the music, which I referto as harmonic priority; the selection of notes with particular attention to the shape

of the melodic line, which I refer to as melodic priority; and the recall of music played

earlier in the solo, which then is incorporated into the ongoing line.









I describe these processes in detail below. With each description, I provide exam-

ples from individual participants that illustrate each idea.

Sketch Planning

In the process of sketch planning, one or more musical features of upcoming passages

are conceived by the improviser before the passages’ performance. These features include

architectural elements like note density, instrument register, and harmonic structure. In

all instances of sketch planning, the improviser is aware of these features before the

passage is played. Every participant in the study described planning upcoming aspects

of his solo. Across the seven interviews, the artists described a total of 43 instances ofsketch planning.

The trombonist, for example, described knowing that his second solo chorus would

“string more linear stuff together.” During the chorus, he played longer eighth-note

lines, departing from the shorter, simpler rhythmic figure that dominated his first

improvised chorus. The trombonist clearly had a plan for the second chorus, but it

included only incomplete information. According to his comments, note density was

an important aspect of the second chorus and was an essential feature of the architec-

ture of the solo; he planned to use eighth-note lines, but decisions about pitch were not

made ahead of the performance of each phrase.In some instances, improvisers described sketching out upcoming passages in sev-

eral stages that became progressively more detailed with the passage of time. The gui-

tarist, for example, explained how he planned to “stretch the listener’s ear” at one point

in his solo, an effect he accomplished by using pitches from outside the key. During the

interview, he clearly stated that he had not determined the device he would use to

“stretch” at the time he derived his sketch. He identified the point at which he made the

decision to use nonchord tones, which occurred several beats after the decision to

stretch (see Figure 1). The guitarist chose to add harmonic tension by using pitches

derived from the altered scale before deciding which pitches to use. He explained thathe chose which notes to play in the moment while the passage was being performed. In

the excerpt in Figure 1, what began as the sketch became progressively more explicit:

stretch the listener, use a dissonant scale, and select pitches from the scale.

Figure 1. Illustration of sketch planning in stagesNote: Excerpt from the guitarist’s solo, mm. 25–28, with decision points indicated.








Norgaard 117

Evaluative Monitoring

The second ongoing process I identified in the thinking of the participants was an evalu-

ative monitoring process. Improvisers described monitoring and evaluating their own

playing as they performed and using that information in subsequent decision making.

I identified 23 instances of the monitoring process in five of the seven transcriptions.

Interestingly, several participants indicated that this monitoring revealed notes that

were “unexpected,” suggesting that improvisers are at times not fully aware of the

sounds they are about to play. The violinist, for example, explained how he monitored

his playing to identify interesting material: “The hand is gonna crawl around . . . and

the brain is gonna like try to pick out something that the hand is doing.” The pianistdescribed a similar process: “Hopefully something pops up that’s worth doing some-

thing with.” In all cases, it appears that the process the improvisers described is sepa-

rate from the conscious control of physical movements. When the violinist describes

the hand as “crawling around” and “the brain” as picking out material, he gives the

impression that the brain is monitoring what the hand is doing and the hand is not

under conscious control.

In another example, the bassist referred to a melodic variation by saying, “I hear

myself doing that.” His comment concerned the improvised bass line in the first mea-

sure of his second chorus, where the pitches do not appear to match the underlyingF chord (see Figure 2). The figure is based on the F triad, but the placement is shifted

rhythmically. The figure starts with a chromatic approach note on Beat 1, and the root, F,

does not appear until the third beat. This creates melodic tension that he resolves five

measures later on the Bb chord by placing the root on the first beat. The bassist’s com-

ment implies that this effect was not planned but instead occurred without deliberate

thought and that he was reacting to the sounds he heard coming from his own instru-

ment, noting how his own line had added tension and therefore needed to “be resolved.”

The lack of deliberative control sometimes results in improvisational outcomes that

improvisers label as errors. In one example, the bassist pointed out how an improvisedline ended unexpectedly. “I got there too soon or sooner than I would have liked, so

you have to be a helicopter and kind of hover around. I had to fix that, whatever that is.”

The unexpected occurrence had to be accommodated by altering subsequent

Figure 2. Illustration of monitoringNote: Excerpt from the bassist’s solo, mm. 12–17, with relevant comment from his interview. Thedissonant chromatic approach note in m. 13 and the root on the first beat of m. 17 are indicated.









improvised material. The bassist simply incorporated the end of the melodic line that

ended unexpectedly into the following phrase. The result is improvised material that

appears perfectly coherent, even though it derived from an unexpected event.

Four Generative Strategies

The two processes described above represent thinking that appears to be ongoing through-out the creation of each improvisation, but the processes do not describe how musical

material is initially conceived. From the participant interviews, I identified four gen-

erative strategies that describe ways of creating improvised material—making choices

about which notes to play. Improvisers seemed to switch from one generative strategy

to another at different times during their solos, and the amount of material created using

each generative strategy differed among participants. Using material from an idea bank

is a flexible way of incorporating practiced musical elements into a solo. I define the

idea bank as the collection of procedural and auditory memories of coherent musical

structures, which may vary in explicitness and extent. I found 76 instances in which participants described drawing material from an idea bank.

Some ideas from the idea bank are inserted as they are recalled, with no modifica-

tions from the memory stored. The pianist, for example, described inserting complete

remembered ideas in parts of his solo (see Figure 3). He labeled these figures “units,”

which he can replicate exactly in multiple contexts. Regarding the excerpt, the pianist

described his improvisational thinking as connecting smaller units to form longer lines.

He compared this to building with Legos, in which creation is a process of connecting

pre-formed blocks.

Several participants mentioned using melodic figures that were unique versions oflearned models, accessing a learned melodic figure from long-term memory and

adapting it to fit the current context. Some participants referred to this phenomenon as

using a model. The model may have all the information needed to perform an actual

Figure 3. Illustration shows explicit ideas or “units” from the idea bank

Note: The units are marked with brackets. Excerpt from the pianist’s solo, mm. 57–60, with relevant com-ments from his interview.








Norgaard 119

melodic figure, but the context requires the improviser to modify the version stored in

memory. The trombonist, for example, described how he learned to play over a par-

ticular set of chords in the seventh and eighth bars of the blues by memorizing a model played by Kenny Dorham that uses an advanced harmonic device referred to as a tri-

tone substitution. The trombonist described a melodic figure in his solo as an adapta-

tion of the model as played by Dorham (see Figure 4).

In some cases, the ideas recalled from the idea bank are only templates or outlines.

The alto saxophonist, for example, explained how he constructed a melodic figure

using a memorized melodic contour as a template. “I’m using a rising scale and then

when the five chord hits that diminished triad that we map on to the dominant chord . . .

it is a general shape . . . a scale orientation of the two chord and an arpeggiated version

of the five chord.” He asserted that the version of the template in this instance wasunique and he had not played it previously.

The improvisers sometimes referred to vague stylistic features as guiding their cre-

ation of musical figures. These ideas from the idea bank are inexplicit memories of

sounds or effects. In one instance, for example, the trombonist described how a melodic

figure ends “the way Parker ends phrases.” Yet, when asked to define these features,

he referred only to inexplicit labels like “cool” and “laid back.”

Harmonic priority is the generative strategy defined by selecting pitches with atten-

tion focused on the music’s chordal structures. When improvising with harmonic

priority, improvisers create melodic lines by shaping melodic material to fit thechord structure. Participants who described improvising with harmonic priority identi-

fied “target notes” among available chord tones and created melodic passages that

placed these targets on the strong beats of the measure. I identified 91 musical exam-

ples illustrating the use of harmonic priority as a generative strategy.

Figure 5 shows a long eighth-note line that, according to the guitarist, was created

using harmonic priority as a generative strategy. In the example, all the chord tones are

circled, and the chord tones that appear on Beats 1 and 3 are marked with arrows. The

example illustrates the prevalence of chord tones on strong beats. The few measures

that feature non chord tones on the strong beats introduce tension that is resolved insubsequent measures with chord tones again placed on the strong beats. The guitarist

described his thinking process as “weaving through the changes.”

Figure 4. Illustration shows a figure adapted from a modelNote: Excerpt from the trombonist’s solo, mm. 54–57, with relevant comments from his interview.









When artists improvise with melodic priority, they devote less attention to the

underlying chord progression and focus more on the horizontal features of the impro-

vised material. This often results in melodies based on a single scale that span several

chord changes. This approach is described by one participant as “blanketing.”

When discussing the use of melodic priority, participants often mentioned disre-

garding the harmonic underpinnings of the blues progression. The trombonist describedthis linear process as “singing” and indicated that he did not consciously consider

chord information as he played (see Figure 6). Fewer musical examples of the melodic

priority generative strategy were identified in the transcriptions compared to the other

strategies. Nine examples were identified in three of the seven transcriptions.

The fourth generative strategy involves repeating material played earlier in the solo,

either exactly or with modifications. Several improvised solos included examples

where melodic figures are repeated with minor modifications. In one example, the pia-

nist repeated a short rhythmic figure four times with alterations that implied the under-

lying chord progression. I identified 27 examples in which the improviser repeatedmaterial played earlier in the same solo.

The violinist spoke about relating his entire solo to a short melodic theme that

he identified in the second solo chorus. Throughout the rest of the solo, he pointed out

Figure 6. Illustration of melodic priorityNote: Excerpt from the trombonist’s solo, mm. 92–95, with relevant comment from his interview.

Figure 5. Illustration shows an eighth-note line created with the harmonic priority

generative strategyNote: The circled notes denote chord tones and the arrows point to chord tones on Beats 1 and 3.Excerpt from the guitarist’s solo, mm. 29–36. Note that the F major chord in m. 35 is anticipated in theend of m. 34.








Norgaard 121

instances where he incorporated this theme. The excerpt in Figure 7 indicates where he

identified the melodic figure that became the theme for the rest of his solo. Figure 8 illus-

trates how this figure is used several choruses later. It appears that the decision to

incorporate the figure later in the improvisation was made after the figure was per-

formed. This suggests that the evaluative monitoring process mentioned previously plays a role in decisions about repeating material in the solo.

Discussion

I identified six main themes that capture the various thought processes engaged by

expert improvisers. I obtained the results through retrospective verbal reports in which

the improvisers described their own thinking as they viewed notation and listened to

their improvisations minutes after they had been performed. The procedures allowed

me to link the participants’ verbal comments to specific musical material. This inves-tigation represents the first time that improvisers answered questions about their

behavior while looking at notation of their just-completed performances. It should be

acknowledged that the participants’ descriptions were based on reflections about their

own thinking and may not fully reflect their thoughts during their improvisations.

There are intriguing parallels between published models of improvisation and themes

identified in the current study. Two theoretical models mentioned earlier describe how

decisions may occur either at specific time points between note groupings (Pressing,

1988) or continuously as notes are chosen implicitly according to internalized rules

(Johnson-Laird, 2002). The current study suggests that improvisers may use both pro-cesses at different times.

Pressing’s (1988) observation that notes are chosen in groups is consistent with

participants’ descriptions of drawing intact motives from a memorized idea bank. In

Figure 7. Illustration shows where the “theme” of the solo (circled) is identifiedNote: Excerpt from the violinist’s solo, mm. 37–40, with relevant comment from his interview.

Figure 8. Illustration shows later use of the thematic figureNote: Excerpt from the violinist’s solo, mm. 93–96, with relevant comment from his interview.









most instances observed in these improvisations, ideas were not inserted note for note

but were adapted in accordance with the harmonic, melodic, and stylistic contexts in

which they were applied. The decision to insert a group of notes into a solo line sets in

motion the recall of a procedural memory that plays out until the note grouping iscompleted. Pressing labeled these decision points “triggers” and described how audi-

tory and proprioceptive feedback loops are used to shape subsequent note groupings at

each successive trigger. This is similar to the current participants’ descriptions of mon-

itoring the outcome and using the information to make sketch plans for upcoming

material. A related observation was made by Berkowitz (2010), who explained that “as

the execution of the musical goal takes place, the performer must monitor the result

through both aural and kinesthetic feedback” (p. 147). In addition, Pressing pointed

out that improvisers sometimes repeat ideas in various forms just as current partici-

pants described using material repeatedly throughout their solos.The harmonic and melodic priority generative strategies are consistent with

Johnson-Laird’s (2002) theory of internalized implicit rules as the main guiding force

behind note choices. Participants in the current study described creating longer lines

that aligned with the chord progression without focusing on individual note choices.

For example, the pianist described the thought behind a line that clearly reflected the

underlying harmonic structure as being concerned with the contour of the line: “I think

what I think about there is like ‘How can I gracefully get down from this register?’ you

know, and then secondarily to that, okay, I’m gonna play notes that reflect the chords of

the blues.” This implies that selecting notes that fit the harmonic progression requireslittle effort and that the improviser is focused mainly on shaping the melodic contour.

Similarly, using melodic priority, the bassist described, “I’m following the shape of

my line. I don’t really know where it’s gonna take me.” Internalized melodic rules appear

to shape the line continuously outside of conscious control while the improviser moni-

tors the outcome.

The processing speed necessary to create improvisations in real time also suggests

that individual notes are not chosen consciously (Dietrich, 2004; Johnson-Laird, 2002;

Kenny & Gellrich, 2002; Pressing, 1988), and participants in the current study described

feelings of automaticity during their improvisation. Berkowitz (2010) referred to the phenomenon as “creator and witness” in which the improviser creates notes with a fast

implicit process that is witnessed by the conscious mind only after the notes are played.

This notion is supported by brain imaging data that suggests that the part of the brain

that is active during conscious decision making is less engaged during improvisation

than during the performance of learned material (Limb & Braun, 2008).

Although individual notes may be chosen implicitly, the current study suggests that

those choices may be guided consciously by the overall goal of creating architecturally

interesting improvisations. The pianist described how he often starts a blues improvi-

sation by visualizing a “solo curve”: “Start the solo with sparse melodic material,develop this material in subsequent choruses building to an emotional peak in the second

to last chorus, and finally ‘wind down’ in the last chorus.” Participants in the Limb and

Braun (2008) study listened to a prerecorded accompaniment and were not asked to







Norgaard 123

create improvisations with an architectural structure, possibly accounting for the deac-

tivation in the brain regions associated with conscious decision making.

In an actual jazz performance, the improviser may be engaged primarily in interac-

tion with other ensemble members (Berliner, 1994; Monson, 1996; Sawyer, 1992).Most of the comments in the study of jazz duos by Mendonça and Wallace (2004) con-

cerned how interaction affected decisions. Interaction was eliminated as a factor both

in the current study and in the study by Limb and Braun (2008). Further research should

explore the effects of larger architectural plans and interactivity on improvisational

behavior and thinking.

Implications for Music Education

It has been argued that traditional jazz instruction emphasizes music theory at theexpense of structural and interactive concerns (Kenny & Gellrich, 2002; Pressing,

1988; Sarath, 2002). As a result, novice jazz improvisers may develop thinking patterns

that focus on individual note choices in ways that are unlike the thinking of skillful

improvisers. Although knowledge of jazz theory may be common among jazz artists,

so are skills related to compositional architecture and interaction among performers.

With this in mind, it may be that aspiring jazz musicians would benefit from work-

ing among several different modes of thinking. For example, (1) a “theory mode” that

explores the idea bank and harmonic priority generative strategies, with conscious

attention focused on technical and theoretical concepts, and (2) a “play mode” thatfocuses on planning and evaluative processes in addition to interaction with other stu-

dents and teachers.

Instruction in theory mode could include the deliberate building of the idea bank

and procedures for improvisation using harmonic priority. Traditional jazz pedagogi-

cal materials often are designed to develop the idea bank and teach students to empha-

size chord tones in their improvisations (e.g., Baker, 1988; Coker, 1980). Techniques

for learning to use prelearned fragments include deliberately inserting a particular

figure in the improvisation at a spot that fits the chordal structure, such as the trombon-

ist described in the current study using a learned Kenny Dorham figure. The goal ofinstruction in theory mode should be that the student through practice will learn to use

these techniques without conscious effort.

Instruction in play mode could cultivate the ability to improvise coherently by

explicitly structuring student learning activities that do not require conscious attention

to the chord structure. Activities in play mode could include planning and executing

longer solo structures using effects such as note density and register similar to the plan

described by the pianist in the current study. Central to the idea of play mode is that the

chordal structure accompanying the improvisational activities should be so familiar to

the student that all attention can be focused on planning, monitoring, and interactive processes.

The results of the current study also could be used to design meaningful introduc-

tory improvisation exercises for band and orchestra students. This study’s results








suggest that experts are concerned primarily with planning and monitoring their own

output in order to create structurally sound improvisations. Actual note choices are

guided by practiced implicit processes that rely on a bank of ideas and procedures for

creating melodically and harmonically appropriate musical material. Students whoengage in introductory improvisation exercises do not have such a bank of material

and processes to rely on. However, simply growing up in the Western musical culture

appears to support basic improvisation without prior improvisational instruction

(Baldi, Tafuri, & Caterina, 2002; Brophy, 2005; Kiehn, 2003; Kiehn & Bay, 2007).

Some of the improvisation exercises used in this research with young students could

be transferred to the band or orchestra classroom. Brophy (2005) constructed a class

setting in which all students took turns improvising on a xylophone using a scale in

which all available notes fit the accompaniment. Similarly, a diatonic chordal accom-

paniment could be used in exercises that provide orchestra or band students an oppor-tunity to improvise individually using a familiar scale. This would allow instruction to

be focused on the use of motives and the development of phrases and to demonstrate

how improvisers can monitor their output to choose ideas for further development.

The primary goal of improvisation instruction should be to structure learning activ-

ities in which students at all levels can experience ways of thinking that resemble those

of artist-level improvisers. Accomplishing this goal requires that we more closely

examine and more clearly describe not only what experts do but also how they think

about what they do. The research described in this report provides a new perspective

on improvisational thinking and may help illuminate the covert planning and decisionmaking that lead to artistic creation in jazz.

Author’s Note

This article is based on the author’s dissertation, “Descriptions of Improvisational Thinking by

Artist-Level Jazz Musicians,” completed at The University of Texas at Austin in 2008.

Declaration of Conflicting Interests

The author declared no potential conflicts of interests with respect to the research, authorship,

and/or publication of this article.

Funding

The author received no financial support for the research, authorship, and/or publication of this

article.

References

Azzara, C. D. (2002). Improvisation. In R. Colwell & C. P. Richardson (Eds.), The new hand-

book of research on music teaching and learning: A project of the Music Educators National

Conference (pp. 171–187). Oxford England: Oxford University Press.Baker, D. N. (1988). Jazz improvisation (2nd ed.). Van Nuys, CA: Alfred.








Norgaard 125

Baldi, G., Tafuri, J., & Caterina, R. (2002). The ability of children aged 7-10 to structure musi-

cal improvisations. Bulletin of the Council for Research in Music Education, pp. 135–141.

Berkowitz, A. L. (2010). The improvising mind: Cognition and creativity in the musical moment .

New York, NY: Oxford University Press.Berliner, P. F. (1994). Thinking in jazz . Chicago, IL: University of Chicago Press.

Brophy, T. S. (2005). A longitudinal study of selected characteristics of children’s melodic

improvisations. Journal of Research in Music Education, 53, 120–133.

Bruner, J. (1977). The process of education. Cambridge, MA: Harvard University Press.

Campbell, P. S., Connell, C., & Beegle, A. (2007). Adolescents’ expressed meanings of music in

and out of school. Journal of Research in Music Education, 55, 220–236.

Carman, P. (1999). CD metronome: Vol. 1 straight ahead swing 4/4 [CD]. Idyllwild, CA: Paul

Carman Music.

Charmaz, K. (2006). Constructing grounded theory. London, England: Sage.Clarke, E. F. (1988). Generative principles in music performance. In J. A. Sloboda (Ed.),

Generative processes in music: The psychology of performance, improvisation, and compo-

sition (pp. 1–26). Oxford, England: Clarendon Press.

Coker, J. (1980). Jerry Coker’s complete method for improvisation: For all instruments. Van

Nuys, CA: Alfred.

Dietrich, A. (2004). Neurocognitive mechanisms underlying the experience of flow. Conscious-

ness and Cognition: An International Journal , 13, 746–761.

Duke, R. A. (2005). Intelligent music teaching: Essays on the core principles of effective instruc-

tion. Austin, TX: Learning and Behavior Resources.Egorov, A. (2004). TS-Audio To Midi Realtime Converter (Version 3.30) [Computer software].

TallStick Software.

Finkelman, J. (1997). Charlie Christian and the role of formulas in jazz improvisation.

Jazzforschung/Jazz Research, 29, 159–188.

Glaser, B. G. (1998). Doing grounded theory: Issues and discussions. Mill Valley, CA: Sociology

Press.

Glaser, B. G., & Strauss, A. (1967). The discovery of grounded theory: Strategies for qualitative

research. New York, NY: Aldine de Gruyter.

Gridley, M. C. (1987). Jazz styles: History & analysis. Englewood Cliffs, NJ: Prentice Hall.Gushee, L. (1991). Lester Young’s “Shoe shine boy.” In L. Porter (Ed.), A Lester Young reader

(pp. 224–254). Washington, DC: Smithsonian Institution Press.

Hargreaves, D. J., Cork, C. A., & Setton, T. (1991). Cognitive strategies in jazz improvisation:

An exploratory study. Canadian Journal of Research in Music Education, 33, 47–54.

Harley, T. A. (2008). The psychology of language: From data to theory (3rd ed.). Hove, Eng-

land: Psychology Press.

Henwood, K., & Pidgeon, N. (2003). Grounded theory in psychological research. In P. M. Camic,

J. E. Rhodes, & L. Yardley (Eds.), Qualitative research in psychology: Expanding perspec-

tives in methodology and design (pp. 131–155). Washington, DC: American PsychologicalAssociation.

Johnson-Laird, P. N. (2002). How jazz musicians improvise. Music Perception, 19, 415–442.









Kenny, B. J., & Gellrich, M. (2002). Improvisation. In R. Parncutt & G. E. McPherson (Eds.),

The science and psychology of music performance (pp. 117–134). New York, NY: Oxford

University Press.

Kiehn, M. T. (2003). Development of music creativity among elementary school students. Journalof Research in Music Education, 51, 278–288.

Kiehn, M., & Bay, G. (2007). Creative thinking: Music improvisational skills development

among elementary school students. Education and Human Development , 1(2), 1–13.

Levelt, W.J.M., Roelofs, A., & Meyer, A. S. (1999). A theory of lexical access in speech produc-

tion. Behavioral and Brain Sciences, 22, 1–75.

Limb, C. J., & Braun, A. R. (2008). Neural substrates of spontaneous musical performance: An

fMRI study of jazz improvisation. PLoS ONE , 3(2), e1679.

Mendonça, D., & Wallace, W. A. (2004, August). Cognition in jazz improvisation: An explor-

atory study. Paper presented at the Cognitive Science Society Annual Meeting, Chicago, IL.Monson, I. (1996). Saying something: Jazz improvisation and interaction. Chicago: University

of Chicago Press.

Muhr, T. (2006). ATLAS.ti (Version 5.2.18) [Computer software]. Berlin, Germany: ATLAS.ti

Scientific Software.

Owens, T. (1995). Bebop: The musicians and its players. New York, NY: Oxford University

Press.

Patel, A. D. (2003). Language, music, syntax and the brain. Nature Neuroscience, 6 , 674–681.

Pressing, J. (1984). Cognitive processes in improvisation. In W. R. Crozier & A. J. Chapman

(Eds.), Cognitive processes in the perception of art (pp. 345–363). Amsterdam, the Nether-lands: Elsevier Science.

Pressing, J. (1988). Improvisation: Methods and model. In J. A. Sloboda (Ed.), Generative pro-

cesses in music (Paperback 2000 ed., pp. 129–178). Oxford, England: Oxford University

Press.

Pressing, J. (1998). Psychological constraints on improvisational expertise and communication.

In B. Nettl & M. Russell (Eds.), In the course of performance: Studies in the world of musi-

cal improvisation (pp. 47–68). Chicago, IL: University of Chicago Press.

Rubin, H. J., & Rubin, I. S. (2005). Qualitative interviewing: The art of hearing data (2nd ed.).

Thousand Oaks, CA: Sage.Sarath, E. (2002). Improvisation and curriculum reform. In R. Colwell & C. P. Richardson

(Eds.), The new handbook of research on music teaching and learning: A project of the

Music Educators National Conference (pp. 188–198). Oxford, England: Oxford University

Press.

Sawyer, R. K. (1992). Improvisational creativity: An analysis of jazz performance. Creativity

Research Journal , 5, 253–263.

Schuller, G. (1968). Early jazz: Its roots and musical development . New York, NY: Oxford

University Press.

Strauss, A., & Corbin, J. (1998). Basics of qualitative research: Techniques and procedures fordeveloping grounded theory (2nd ed.). Thousand Oaks, CA: Sage.

Tost, T. (2007). Samplitude Professional (Version 9.1) [Computer software]. Lübbecke, Germany:

Magix AG.








Norgaard 127

Willis, J. W. (2007). Foundations of qualitative research: Interpretive and critical approaches.

Thousand Oaks, CA: Sage.

Bio

Martin Norgaard is an assistant professor of music education at Georgia State University. His

research interests include the integration of creative activities in the classroom, string education,

and improvisation in all its forms.

Submitted October 9, 2009; accepted October 12, 2010.

Documents

Journal of Research in Music Education-2011-Norgaard-109-27